Tools for Feature Extraction Exploring and using the essentia - - PowerPoint PPT Presentation

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Tools for Feature Extraction

Exploring and using the essentia library

• A. Hanke1
• A. Paranjape2

1andrea.hankerwth-aachen.de 2akshay.paranjaperwth-aachen.de

Topics in Computer Music, RWTH Aachen, July 2017

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 1/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Outline

1 Introduction to essentia

General structure of essentia How To: Use essentia with Python How To: Add a new function in C++

2 Danceability()

Description of the Danceability-algorithm Modifying the algorithm Results and interpretation of the modified algorithm

3 Classifying Songs

General idea of the proposed algorithm ”Sneak Peak” into the extraction algorithm

4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 2/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Section 1

1 Introduction to essentia

General structure of essentia How To: Use essentia with Python How To: Add a new function in C++

2 Danceability() 3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 3/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Section 1

1 Introduction to essentia

General structure of essentia How To: Use essentia with Python How To: Add a new function in C++

2 Danceability() 3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 4/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

A few remarks and little bit of history

Essentia is a C++ library for audio analysis with Python bindings

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 5/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

A few remarks and little bit of history

Essentia is a C++ library for audio analysis with Python bindings fairly young (first version released in 2008, newest version 2.1 beta3 released on September 2016)

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 5/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

A few remarks and little bit of history

Essentia is a C++ library for audio analysis with Python bindings fairly young (first version released in 2008, newest version 2.1 beta3 released on September 2016)

• pen-source, with continuing additions

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 5/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation statistics and math

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation statistics and math tonal and pich analysis

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation statistics and math tonal and pich analysis duration, silence

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation statistics and math tonal and pich analysis duration, silence loudness, dynamics and rythm

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Some features of essentia

Essentia provides algorithms for spectral analysis extraction and segmentation statistics and math tonal and pich analysis duration, silence loudness, dynamics and rythm filters ...

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 6/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Operating modes

Essentia provides two operating modes: standard, recommended for

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 7/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Operating modes

Essentia provides two operating modes: standard, recommended for

maximum control in C++ research with Python (interactive environment)

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 7/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Operating modes

Essentia provides two operating modes: standard, recommended for

maximum control in C++ research with Python (interactive environment)

streaming, recommended for

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 7/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General structure of essentia

Operating modes

Essentia provides two operating modes: standard, recommended for

maximum control in C++ research with Python (interactive environment)

streaming, recommended for

easy-to-write extractors in C++ and Python porting from Python to C++

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 7/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Section 2

1 Introduction to essentia

General structure of essentia How To: Use essentia with Python How To: Add a new function in C++

2 Danceability() 3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 8/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Using essentia in Python

Step 0: Import the library import e s s e n t i a import e s s e n t i a . standard import e s s e n t i a . streaming

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 9/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Using essentia in Python

Step 0: Import the library import e s s e n t i a import e s s e n t i a . standard import e s s e n t i a . streaming Step 1: Instantiate algorithms from the library l o a d e r = e s s e n t i a . standard . EasyLoader ( filename = m u s i c f i l e , startTime = musicStart , endTime = musicEnd )

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 9/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Using essentia in Python

Step 0: Import the library import e s s e n t i a import e s s e n t i a . standard import e s s e n t i a . streaming Step 1: Instantiate algorithms from the library l o a d e r = e s s e n t i a . standard . EasyLoader ( filename = m u s i c f i l e , startTime = musicStart , endTime = musicEnd ) Step 2: Use the algorithms audio = l o a d e r ()

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 9/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Algorithms in essentia

can have multiple parameters, set during the instantiation

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 10/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Algorithms in essentia

can have multiple parameters, set during the instantiation inputs, set when using an algorithm

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 10/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Use essentia with Python

Algorithms in essentia

can have multiple parameters, set during the instantiation inputs, set when using an algorithm

• utputs

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 10/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Section 3

1 Introduction to essentia

General structure of essentia How To: Use essentia with Python How To: Add a new function in C++

2 Danceability() 3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 11/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 1:

Create the .h and .cpp files: danceabilityDetailed.h danceabilityDetailed.cpp

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 12/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

start of file, class declaration #ifndef ESSENTIA DANCEABILITYDETAILED H #define ESSENTIA DANCEABILITYDETAILED H #include ” algorithm . h” #include ” essentiamath . h” namespace e s s e n t i a { namespace standard { class D a n c e a b i l i t y D e t a i l e d : public Algorithm {

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 13/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

declare all protected variables, especially input and output: protected : Input<std : : vector <Real> > s i g n a l ; Output<Real> d a n c e a b i l i t y ; Output<std : : vector <Real> > dfaExponents ; Output<std : : vector <Real> > dfaTaus ; int p r e f e r r e d S i z e , a c t u a l S i z e ;

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 14/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

constructor declaration, here input and output are declared public : D a n c e a b i l i t y D e t a i l e d () { p r e f e r r e d S i z e = 36; // as seen in the paper a c t u a l S i z e = p r e f e r r e d S i z e ; d e c l a r e I n p u t ( s i g n a l , ” s i g n a l ” , ”#d” ) ; declareOutput ( dfaTaus , ” dfaTaus ” , ”#d” ) ; }

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 15/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

declare parameters void declareParameters () { declareParameter ( ”minTau” , ”#d” , ” (0 , i n f ) ” , 3 1 0 . ) ; declareParameter ( ”maxTau” , ”#d” , ” (0 , i n f ) ” , 8 8 0 0 . ) ; }

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 16/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

declare other functions or variables void compute ( ) ; void c o n f i g u r e ( ) ; s t a t i c const char∗ name ; s t a t i c const char∗ category ; s t a t i c const char∗ d e s c r i p t i o n ;

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 17/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

declare other functions or variables protected : std : : vector <int> tau ; Real stddev ( const std : : v ec t or <Real>& array , int s t a r t , int end ) const ; }; } // namespace standard } // namespace e s s e n t i a

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 18/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 2: danceabilityDetailed.h

declare other functions or variables protected : std : : vector <int> tau ; Real stddev ( const std : : v ec t or <Real>& array , int s t a r t , int end ) const ; }; } // namespace standard } // namespace e s s e n t i a Repeat for streaming...

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 18/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 3: danceabilityDetailed.cpp

File start and fill inherited variables #i n c l u d e ” d a n c e a b i l i t y D e t a i l e d . h” using namespace std ; namespace e s s e n t i a { namespace standard { const char∗ D a n c e a b i l i t y D e t a i l e d : : name= ” D a n c e a b i l i t y D e t a i l e d ” ;

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 19/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 3: danceabilityDetailed.cpp

File start and fill inherited variables const char∗ D a n c e a b i l i t y D e t a i l e d : : category = ”Rhythm” ; const char∗ D a n c e a b i l i t y D e t a i l e d : : d e s c r i p t i o n = DOC( ”Long\n\ n D e s c r i p t i o n ” ) ;

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 20/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary How To: Add a new function in C++

Step 3: danceabilityDetailed.cpp

Fill the declared functions with meaning Real D a n c e a b i l i t y D e t a i l e d : : stddev ( const vector <Real>& array , int s t a r t , int end ) const { . . . } void D a n c e a b i l i t y D e t a i l e d : : c o n f i g u r e () { . . . } void D a n c e a b i l i t y D e t a i l e d : : compute () { . . . }

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 21/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Section 2

1 Introduction to essentia 2 Danceability()

Description of the Danceability-algorithm Modifying the algorithm Results and interpretation of the modified algorithm

3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 22/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Description of the Danceability-algorithm

Section 1

1 Introduction to essentia 2 Danceability()

Description of the Danceability-algorithm Modifying the algorithm Results and interpretation of the modified algorithm

3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 23/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Description of the Danceability-algorithm

Sample frame title

Detrended Fluctuation Function - statistical self affinity y(m) = m

n=1 s(n)

Computation procedure

Segmentation Standard Deviation

Danceability taken as the average of the DFT Higher the Danceability value, more is the music danceable

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 24/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Modifying the algorithm

Section 2

1 Introduction to essentia 2 Danceability()

Description of the Danceability-algorithm Modifying the algorithm Results and interpretation of the modified algorithm

3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 25/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Modifying the algorithm

Modifying Essentia algorithm

Essentia provides open source C++ library with Python interface Danceability() function

Input - signal(vector real) Output - danceability, DFA(real)

Modified the code as according to get vector as output Danceability values for Music

ChaChaCha - 1.18 Rumba - 1.05 Tango - 1.28

Inference - Tango is more Danceable

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 26/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Results and interpretation of the modified algorithm

Section 3

1 Introduction to essentia 2 Danceability()

Description of the Danceability-algorithm Modifying the algorithm Results and interpretation of the modified algorithm

3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 27/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Results and interpretation of the modified algorithm

Results

Figure: From the research paper: time series s(n)

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 28/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Results and interpretation of the modified algorithm

Results

Figure: From the research paper: DFA Exponent Function

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 29/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary Results and interpretation of the modified algorithm

Let’s take a look and try it out!

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 30/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Section 3

1 Introduction to essentia 2 Danceability() 3 Classifying Songs

General idea of the proposed algorithm ”Sneak Peak” into the extraction algorithm

4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 31/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General idea of the proposed algorithm

Section 1

1 Introduction to essentia 2 Danceability() 3 Classifying Songs

General idea of the proposed algorithm ”Sneak Peak” into the extraction algorithm

4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 32/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General idea of the proposed algorithm

Purpose of the algorithm

is to extract features from songs

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 33/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General idea of the proposed algorithm

Purpose of the algorithm

is to extract features from songs

beats per minute meter danceability ...

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 33/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General idea of the proposed algorithm

Purpose of the algorithm

is to extract features from songs

beats per minute meter danceability ...

classify them according to their features (possibly with machine learning)

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 33/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary General idea of the proposed algorithm

Purpose of the algorithm

is to extract features from songs

beats per minute meter danceability ...

classify them according to their features (possibly with machine learning) be used as classification for DJ-algorithm

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 33/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Section 2

1 Introduction to essentia 2 Danceability() 3 Classifying Songs

General idea of the proposed algorithm ”Sneak Peak” into the extraction algorithm

4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 34/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Extraction demonstration of

beats per minute

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 35/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Extraction demonstration of

beats per minute rubato

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 35/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Extraction demonstration of

beats per minute rubato meter

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 35/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Extraction demonstration of

beats per minute rubato meter novelty

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 35/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Extraction demonstration of

beats per minute rubato meter novelty danceability

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 35/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary ”Sneak Peak” into the extraction algorithm

Let’s take a look and try it out!

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 36/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Section 4

1 Introduction to essentia 2 Danceability() 3 Classifying Songs 4 Conclusion and Summary

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 37/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

In Python interface which makes it easy for beginners

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

In Python interface which makes it easy for beginners Inbuilt C++ function giving us the opportunity to alter the code

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

In Python interface which makes it easy for beginners Inbuilt C++ function giving us the opportunity to alter the code Essentia Library is being used world wide for Music Classification based on tempo, rubato etc.

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

In Python interface which makes it easy for beginners Inbuilt C++ function giving us the opportunity to alter the code Essentia Library is being used world wide for Music Classification based on tempo, rubato etc. Danceability approach mentioned by us is easy and well defined to get a quick overview of any song

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38

Introduction to essentia Danceability() Classifying Songs Conclusion and Summary

Conclusion and Summary

Essentia is a powerful tool to extract music features with many

• ptions and algorithms ranging from Spectral, tonal, Rhythm

In Python interface which makes it easy for beginners Inbuilt C++ function giving us the opportunity to alter the code Essentia Library is being used world wide for Music Classification based on tempo, rubato etc. Danceability approach mentioned by us is easy and well defined to get a quick overview of any song Overall we recommend this library for music lovers and who wish to play around, especially rhythm

Andrea Hanke, Akshay Paranjape RWTH Tools for Feature Extraction: Exploring essentia 38/ 38